관계를 통한 엔티티의 이해

Abstract

Understanding entity is important to solve many entity-related problems such as information retrieval, recommendation, and question answering. To help machines understand entities, knowledge bases have been constructed. However, these knowledge bases suffer from the incomplete problem. To solve this incomplete problem, we extract and add three interesting relationship types: 1) social, 2) comparable, and 3) verb relationships. For each relationship type, we propose a solution to extract relationships from external resource by utilizing knowledge bases. We also show their effectiveness empirically.